Structuralism

You know that feeling you have when something niggles away at the back of your mind and won’t go away? Well, for the last 45 years…

It’s not quite that bad. But I was reading a little while ago about Haeckel’s forged drawings of embryos, purportedly showing their recapitulation of evolution. And the writer mentioned the interesting phenomenon that vertebrates which differ markedly in life do indeed go through a stage in which they are, at least to some extent, fairly similar. However, it is found that they differ greatly in the earlier stages of development – a fact which is not only hard to square with Haeckel’s recapitulation theory, but with Darwinian processes themselves.

Not only is it odd for creatures related by descent to have such varied development, but changes to such early and crucial processes would be supposed (from universal experimental experience) inevitably to end in death, not evolution.

This reading stirred a memory from 1968, when I studied embryology in A-level zoology in school. We studied, according to my memory and my miraculously preserved notes, the development of the frog, the primitive chordate Amphioxus, the chicken and the mouse. Quite a nice taxonomic spread, and done from books, prepared microscope slides and actual dissection. Some of my drawings are quite good, though I says it myself as shouldn’t. At least they’re more faithful than Haeckel’s.

To an extent these creatures form an evolutionary sequence, and I can remember at the time being quite surprised that the details of their earliest development turned out to be markedly different – particularly, I remember, with regard to the formation of the coelom (or body cavity), perhaps because it was my first encounter with that structure. I was also surprised that neither my colleagues or teachers thought the disparity a matter of any note. If they were evolutionarily related, I thought, surely the basis of their construction ought to be the same? It was like X-raying two near identical pictures, supposedly by the same old master, and finding one based on a careful ink underdrawing and the other a roughly-painted tempera wash.

Now I’ve just got round to reading this article by Michael Denton on structuralism, in contrast to standard evolutionary functionalism. The examples he gives bring my zoology course back to front-of-memory, as he describes how homologous organisms like varieties of wasp can get to the same body plan by entirely different developmental routes. How can that be?

Structuralism bypasses that, and the awkward difficulty I saw in my embryology, because unlike Darwinism it makes the final form a teleological goal (in the Aristotelian sense), and the development process merely the means to that end, rather than the be-all-and-end-all as it must be in Darwinian evolution. An undirected process has to follow the instruction book to the letter. But someone with a concept of the final goal can take various routes to get there.

Denton says a lot about the pentadactyl limb, which is very topical as it came up at the start of the truncated conversation I had with James Stump on BioLogos recently (we’re still on hold regarding that discussion, by the way). Functionalism says that the vertebrate limb pattern may have had a selective advantage once (though it can give no account of it whatsoever), but has persisted for no very good reason for 400 million years in every vertebrate group from ichthyosaurs to moles to swifts – and in both their fore and hind limbs, too. And despite the fact that several single mutations can produce polydactyly as an aberration. Stump, assuming this fuctionalist model, sees vertebrate (and so human) pentadactyly as “contingent”, indeed as evidence of contingency. (What an odd argument to come from a philosopher, by the way – it takes the form, “All As have property B; therefore B is a contingent property of A.”)

Structuralism says that such homologies arose according to laws of some sort, and persist because those laws still apply across groups of organisms. Those laws might be something like platonist forms in the mind of God, but actually nearly all structuralists in the scientific age have seen them as the outworking of physical and chemical principles.

The problem is, of course, that these are what are called “emergent properties”, and such laws have so far evaded attempts to pin them down. Hence I’ve always been skeptical of emergence theories. But they are no less explicable than how Darwinian processes might dictate stable homologies like pentadactyly, and in general terms they explain more about biological form than any Darwinian mechanism does.

For example, if there were some kind of emergent structural principle of this type in protein evolution, then potentially the combinatorial problems that makes random mutation such a laughable excuse for an explanation for functional proteins is ameliorated. Potentially.

How does this fit with a theistic worldview? Well, if you’ve read any Denton, you’ll realise that the world in which he understands such remarkable emergent laws might appear is a very special one indeed. In Nature’s Destiny he details how, at every level from top to bottom, nature is fine-tuned to favour not only life, but intelligent life like ours. It’s clear that he believes such specific order could not have arisen apart from an extremely intelligent creator.

Neither does he see structuralism as a “theory of everything”. He follows the biological pioneer Richard Owen in proposing a structuralist “core” to living forms, and a more superficial adaptational “surface”: hence the astonishingly apt physical structure of DNA would an emergent, structural,  property of physics and chemistry, but the varying sequences of the genome an historical, functional, process. Both, especially in combination, scream out “design” at some very fundamental level, and both leave plenty of space for divine supervision and direction of the process – there is still a massive amount of specific information to explain.

One last point. If there are indeed so-far unexplained structural laws involved, along with adaptive mechanisms, in biology, then it becomes a moot point how far they extend. Homologies like pentadactyly are routinely used as evidence for common descent – but why should that necessarily be so? They might merely indicate common laws operating in comparable situations. And if that’s the case for limb patterns, why not for homologies at the microscopic level – for genes and proteins? Can we really claim we are tracing common descent by similarities in molecules – especially when, as per my previous post, non-homologies also abound in the data.

But that, of course, like structuralism itself, is an heretical thought. But it’s one that’s bugged me since I was a teenager.